The work

You will turn real maintenance work from a mature, widely used open-source repository into challenging features, refactors and bug fixes for AI training. Day to day, that includes writing reference solutions, evaluating AI-generated code and tests, checking grading rubrics, running the project locally with Docker, and explaining technical shortcomings in your own words.

What the screening emphasizes

Mercor will look for a public, verifiable record of sustained contribution through your own GitHub account. You must qualify either as a maintainer named in MAINTAINERS or CODEOWNERS, or through at least 10 pull request reviews in the last 12 months, or as a core contributor with at least 3 substantive merged pull requests in the last 12 months. The relevant repository must be at least two years old, primarily human-written, actively reviewed and used by real users. Expect follow-up on repository architecture, review decisions, testing strategy and edge cases in AI-written changes.

Logistics

This is a 40-hour-per-week W-2 position with Cincinnatus LLC, with work performed on weekdays and on an exclusive basis. During the engagement, you cannot hold another full-time role or participate in another AI training or evaluation project, including another Mercor project. You must be authorized for U.S. W-2 employment, complete required background checks and confidentiality agreements, and use a company-provided device. The listing does not specify whether the role is remote, how much work is asynchronous or the required daily working hours.

Pay band: $150/hr